Date post: | 02-Jun-2018 |
Category: |
Documents |
Upload: | suman-mishra |
View: | 219 times |
Download: | 0 times |
of 21
8/10/2019 Salle, R.D. - Reconsidering Kant, Friedman, Logical Positivism & the Exact Sciences.pdf
1/21
191
*Received September 2001; revised February 2002.
Send requests for reprints to the author, Department of Philosophy, University ofWestern Ontario, London, Ontario N6A 3K7, Canada; e-mail: [email protected].
Philosophy of Science, 69 (June 2002) pp. 191211. 0031-8248/2002/6902-0002$10.00
Copyright 2002 by the Philosophy of Science Association. All rights reserved.
Reconsidering Kant, Friedman, Logical
Positivism, and the Exact Sciences*
Robert DiSalleDepartment of Philosophy
University of Western Ontario
This essay considers the nature of conceptual frameworks in science, and suggests a
reconsideration of the role played by philosophy in radical conceptual change. On
Kuhns view of conceptual conflict, the scientists appeal to philosophical principles is
an obvious symptom of incommensurability; philosophical preferences are merelysub-
jective factors that play a part in the necessarily circular arguments that scientists
offer for their own conceptual commitments. Recent work by Friedman has persua-
sively challenged this view, revealing the roles that philosophical concerns have played
in preparing the way for conceptual change, creating an enlarged conceptual space in
which alternatives to the prevailing framework become intelligible and can be rationally
discussed. If we shift our focus from philosophical themes or preferences to the process
of philosophicalanalysis,however, we can see philosophy in a different and much more
significant historic role: not merely as an external source of general heuristic principles
and new conceptual possibilities, but, at least in the most important revolutionary de-
velopments, as an objective tool of scientific inquiry. I suggest that this approach offerssome insight into the philosophical significance of Newtons and Einsteins revolution-
ary work in physics, and of the interpretation of their work by (respectively) Kant and
the logical positivists. It also offers insight into the connections between modern phi-
losophy of science and some traditional philosophical concerns about the nature of a
priori knowledge.
Michael Friedman has remarked that Kantian thought stands as a
model of fruitful philosophical engagement with the sciences (1992, xii).
Examining this claim, exploring the senses in which, and the extent to
which, it might be true, is not only an illuminating way to consider the
aims and methods of the philosophy of science. It also promises some
insight into the roles that philosophy has played in science, particularly in
8/10/2019 Salle, R.D. - Reconsidering Kant, Friedman, Logical Positivism & the Exact Sciences.pdf
2/21
192
the great conceptual transformations that physics has undergone in the
last three centuries. Friedman himself has, evidently, done much to fulfill
this promise, yielding a deeper appreciation not only of Kants philosophy
of science, but also of the neo-Kantian aspects of logical positivism; he
has rediscovered some insights of Kant and the logical positivists that
more recent accounts of conceptual change have misunderstood, forgot-
ten, or ignored. In doing so he has presented a compelling picture of the
rationality of conceptual change, and of philosophy as a source of rational
motivation for change.
I will argue, however, that the most significant common theme of Kant
and the positivists does not concern the problem of scientific rationality,
as we now understand it. Their special concern was the kind of conceptual
change that constitutes, more than merely a rational choice, an advanceto a deeper level of insight; they sought to show that a novel conception
can represent, not just a more rational way of serving the predictive and
other aims of science, but, more important, a clearer understanding of
something that the existing theory grasps only dimly, or of concepts that
the existing theory has not clearly defined. At a time when even the em-
pirical progress of science is in question, this idea of conceptual progress
may seem difficult to defend. Defending it requires us to reconsider the
role of philosophy in the evolution of sciencenot merely as an external
source of general heuristic principles and new conceptual possibilities, but,
at least in the most important revolutionary developments, as an objective
tool of scientific inquiry. Evidently Kant and the positivists, in differing
ways, saw such a role for philosophy in science. But the limitations inher-
ent in their respective viewpoints prevented them from seeing it in theclearest light, and perhaps discouraged its further exploration. To over-
come their limitations, we need a better understanding of how revolution-
ary concepts can emerge from a critical philosophical engagement with
established beliefs, and in what sense their emergence can be seen as gen-
uine epistemic progress. In such a case, the resulting philosophical insight
is not merely a motivation for the new theory; it is, in essence, the theory
itself. And the kind of scientific progression that results is not linear or
cumulative, in the sense long discredited by Kuhn, but dialectical.
Kuhns notion of paradigm shift has invited many interpretations, more
than one by Kuhn himself. Unquestionably this notion implies that a com-
munity of scientists collectively changes the framework that defines their
metaphysics, their methods, and the nature of their collective social prac-
tice; what seems unclear is to what extent such a change can be regardedas rational. In any case, Kuhn stated quite clearly how he saw the role of
philosophy in this process: when the scientific methods sanctioned by the
existing paradigm fail to decide the great questions, scientists turn to criti-
cal discourse, that is, to the exchange of claims, counterclaims, and
8/10/2019 Salle, R.D. - Reconsidering Kant, Friedman, Logical Positivism & the Exact Sciences.pdf
3/21
8/10/2019 Salle, R.D. - Reconsidering Kant, Friedman, Logical Positivism & the Exact Sciences.pdf
4/21
8/10/2019 Salle, R.D. - Reconsidering Kant, Friedman, Logical Positivism & the Exact Sciences.pdf
5/21
, , 195
2. In more recent literature on the philosophy of space and time (e.g., Earman 1989),the positivists (especially Reichenbach) are roundly dismissed for their technical errorsconcerning the absolute-relational controversy: they did not understand the function
of absolute motion in Newtons theory; they did not understand the true geometricalcontent of general relativity; they did not understand the theoretical continuity of gen-
eral relativity with earlier theories of space and time; they did not understand where to
draw the line between the factual and the conventional in spacetime theories. But thesevery just criticisms do not address the general philosophical question whether conven-
tion plays some role in our knowledge of space and time, and what bearing this mighthave on the traditional ontological issuesa question which is, accordingly, largely
unexamined in the important literature on the philosophy of space and time of the lastfew decades. (Cf. DiSalle 2002b.)
any such differences. That the speed of light is invariant, or the same in
all inertial frames, is an interpretation rather than a logical consequence
of those experimental facts. Therefore, even on Friedmans view, it is in-
appropriate to say that an empirical principle has been elevated to the
status of a constitutive principle, because the constitutive principle is so
distinct, at least in its theoretical implications, from the original empirical
one. It would be more appropriate to say that an interpretive extension
of the empirical law becomes a constitutive principle. This way of describ-
ing it does not make the transition from the one to the other more trans-
parent; on the contrary, it draws attention to the puzzling character of
this transformation in status. It is understandable that Friedman, follow-
ing the logical positivists, would identify this as the point at which an
element of decision or conventionthough, for Friedman, a rationallyand philosophically intelligible decisionmust be implicated. If the facts
dont immediately present themselves in a form upon which a conceptual
framework can be constructed, evidently they can only take on such a
form through an act of our own. This account emphasizes that constitutive
principles have the function of fixing meaning: in addition to the empirical
facts themselves, we require the coordinative definitionthat fixes their re-
lationship to a given mathematical structure, i.e., that fixes the intended
interpretation of an abstract structure.
Friedman is thus reviving an insight associated especially with Hans
Reichenbach, correcting some of its weaknesses instead of overlooking it
as the post-positivist tradition has tended to do.2 To Reichenbach, among
Einsteins fundamental achievements was to have demonstrated the ne-
cessity for metrical coordinative definitions in several places where empir-ical relations had previously been assumed (1957, 15)thus suggesting
the empirical ill-foundedness of Newtonian physics generally, as if its em-
pirical or operational meaning had never been established. As Friedman
has emphasized, the Newtonian picture of space and time was tied to
experience by clear and relatively successful coordinating principles. What
Einstein called into question was the relation between those principles and
8/10/2019 Salle, R.D. - Reconsidering Kant, Friedman, Logical Positivism & the Exact Sciences.pdf
6/21
196
some novel and surprising experimental facts, namely, that electro-
dynamical phenomena appear to exhibit a surprising set of (Lorentzian)
symmetries. Only in this theoretical and experimental context does it make
sense to question the Newtonian account of simultaneity, and to raise the
objection that the physical processes to which the concept of simultaneity
is directly coordinatedespecially the instantaneous propagation of grav-
itational forceare entirely hypothetical. One could have expected the
gravitational definition of simultaneity and the light-signaling method to
be in perfect accord, in the retrospective determination of simultaneity, as
long as the travel-time of light-signals could be taken into account. But
with the discovery that light-signals appear to violate the Galilean rule for
addition of velocities, the gravitational definition now stands on its own,
unsupported by any direct experiment. Einstein saw that in this context,a new coordinative definition was requiredand that in the framework
constituted by this new definition, the Lorentz invariance of electrody-
namics no longer appears to be contradictory, and no longer requires a
hypothetical explanation.
I would also urge, however, a more general philosophical point against
Friedmans formulation, a point that emerges from Einsteins discussion
of electrodynamics, but was first raised in Poincares philosophical reflec-
tions on geometry. Poincare was, evidently, a founder of the notion of
coordinative definition to which Friedman appeals, but his great insight
was not that we designate certain empirical principles as having coordi-
native status. It was, rather, that we are simply mistaken in regarding
certain principles as empirical in the first place. The principle of free mo-
bility, for example, is a kind of definition in disguise, not because wehave elevated it to this status, but because it simply fails to make an em-
pirical claim, or any kind of synthetic claim in the ordinary sense. For the
concepts to which the principle refers are simply not well defined indepen-
dently of the principle itself. As Poincare argued in his celebrated exchange
with Russell on the foundations of geometry, it cannot be an empirical
claim that shape is preserved in rigid motions, because what we mean
by shape is only that which is preserved in certain motions. Russells
(1897) view, that such principles presuppose our acquaintance with the
concepts that occur in them, simply does not survive Poincares conceptual
analysis.
Ironically, Einsteins argument for special relativity directs just this
kind of philosophical criticism against Poincares own view of electrody-
namics. What Einstein recognized around 1905 was not that the principleof the constancy of the velocity of light could beelevated from an empirical
principle into a constitutive principle. He recognized, rather, that the prin-
ciple was not an empirical principle in the first place, and could not be
one. That the speed of light is the same in all inertial frames is a consti-
8/10/2019 Salle, R.D. - Reconsidering Kant, Friedman, Logical Positivism & the Exact Sciences.pdf
7/21
, , 197
tutive principle, not because we grant it that status, but because, so to
speak, it has that status inherently: it is part of the definition of an inertial
framea concept which, as Einstein shows, is not clearly defined inde-
pendently of this principle. This is an aspect of Einsteins introduction of
the light-principle that deserves to be made clearer. The principle ap-
pears to be a straightforward empirical claim, at least if the notion of
inertial frame is taken for granted, as indeed it appears to be in Einsteins
proposal that the same laws of electrodynamics and optics will be valid
for all frames of reference for which the equations of mechanics hold
good ([1905] 1952, 37). To assume this, however, is to assume that we
can begin with an inertial frame and determine, as a matter of empirical
fact, the velocity of light in that frame. Clearly the Michelson-Morley
experiments assumed this much, as did the entire Maxwell-Lorentz theory.The remarkable difference between the Lorentzian perspective and Ein-
steins, then, is not Einsteins reinterpretations of the results of such mea-
surements, but his recognition that such a measurement is, in the contem-
porary situation, impossible. The velocity of light cannot be measured
relative to an inertial frame, because we can no longer assume that we
have an independent way of constructing an inertial frame in advance.
Again, this thought doesnt appear to emerge in the introductory sec-
tion of Einsteins 1905 paper, which already seems to appeal to the notion
of inertial frame. But it emerges directly in the argument of Section I. The
title of this section is already a kind of Socratic irony with respect to the
Lorentzian view: The discussion of procedures for fixing a reference frame
is called the Kinematical Part, as distinct from the Electrodynamical
Part, but we see immediately that electrodynamics is implicated from thestart. For the laying out of a spatial coordinate system, to say nothing of
fixing a standard of time, already requires some means of determining
simultaneous events, and light-signaling proves to be indispensable. From
the Lorentzian perspective, the measurement of the speed of light takes
place against an already well-defined spatial framework. Einsteins anal-
ysis reverses this order: the construction of the spatial frame of reference
requires the assumption of the constancy of the velocity of light. But if
the spatiotemporal framework is not given in advance, then the light prin-
ciple is, ipso facto, not an empirical statement about the behavior of light
in inertial frames; it is by nature a constitutive principle. Einstein arrives
at this principle, then, not by making a decision to grant constitutivestatus
to a known fact. Rather, it is the result of what I have elsewhere charac-
terized as a conceptual analysis, an analysis that discovers, or uncovers,the constitutive principle that is implicit in a given body of theory and
practice (cf. DiSalle 2002b).
What Einsteins analysis shows is that his definition of simultaneity, far
from being an arbitrary new convention, is in fact the definition that we
8/10/2019 Salle, R.D. - Reconsidering Kant, Friedman, Logical Positivism & the Exact Sciences.pdf
8/21
198
3. The present discussion is in sympathy with, and indebted to, that of Torretti, who
also recognizes conceptual criticism as a catalyzer of scientific change (1989, section
2.5). For there can be no question of choosing between two modes of thought if thevery existence of the one issues from a recognition of the conceptual failings of the
other (1989, 44). His account differs from mine, and, I think, more nearly resemblesthat of the logical positivists, in emphasizing the destructive analyses of poorly-defined
concepts, rather than the emergence of new concepts through conceptual analysis. Tor-rettis point is by itself an important one against Kuhn, who seemed to assume that a
usenot only casually in our informal judgments of simultaneous events
(i.e., the events that we see at the same time), but in the setting up of
frames of reference for any kinematical analysis. Until now (1905), how-
ever, we have done so without regarding this definition as playing any
fundamentally constitutive role. This is because, until now, knowing that
light propagates at a finite speed, we supposed that it provides merely a
local criterion of simultaneity, standing in, as it were, for the universal
and invariant criterion that an infinitely fast signal would provide. And
we supposed, as was already noted, that the local criterion could be
brought into correspondence with the invariant one by way of the addition
of velocities, theoretically if not practically. But experiment shows that the
local criterion behaves as if it were an invariant one, failing to distinguish
frames of reference as required by the law of addition. Light-signaling,then, apart from being the only practical means of determining simulta-
neity, turns out to be the only absolute means, i.e., the only one that
does not depend on the perspective of the observer. As Einstein shows in
Section 2 ([1905] 1952, 4143), the relativity of simultaneity is the neces-
sary consequence of adopting this observer-independent criterion.
Calling Einsteins reasoning a conceptual analysis might seem simplis-
tic, recalling the oversimplified view of the logical positivists: that Einstein
had merely applied some empiricist and antimetaphysical standards to
poorly defined conceptions of space and time. But now we can see the
reasoning as a more subtle process, involving not only destructive analysis,
but the construction of something new through a dialectical engagement
with the old; in this sense it can be seen as a genuinely creative process,
not indeed because it invents something unheard of, but because it discernssomething new in the existing theory and fact. That a philosophical per-
spective provided motivation and credibility to special relativity, and
thereby made possible its replacement of the Lorentz theory, therefore
seems an inadequate description of what actually occurred. It is more
accurate to say that special relativityis a philosophical perspective on the
Lorentz theorya clearer perspective, resulting from analysis of what the
Lorentz theory had assumed unreflectingly, namely, the relationships be-
tween rigid bodies (coordinate systems), clocks, and electromagnetic pro-
cesses (Einstein [1905] 1952, 38).3
8/10/2019 Salle, R.D. - Reconsidering Kant, Friedman, Logical Positivism & the Exact Sciences.pdf
9/21
, , 199
given framework is inherently complete and coherent on its own terms, and that it
could only appear inadequate to an adherent of some other frameworkas if its ownadherents were incapable of critical reflection on it (a view that is admirably charitable
to older conceptual frameworks, but, arguably, uncharitable, even condescending, toolder generations of scientists).
It is neither an exaggeration nor a mere metaphor to call Einsteins
critical conceptual analysis dialectical, and in two classical senses of the
term. First, the analysis is a kind of dialogue with an existing point of
view, starting from the latters unexamined assumptions in order to reveal
the weaknesses of its fundamental concepts. In Einsteins Kinematical
Part, the premise of the Lorentzian point of view is provisionally ac-
cepted: that we can begin a physical inquiry by laying down a frame of
referencein Poincares terms, that we can treat the spatio-temporal
background (classical kinematics) as the constitutive framework presup-
posed by the study of any specific field such as electrodynamics. The anal-
ysis quickly reveals, however, that the classical notion of a frame of ref-
erence cannot stand on its own, and that electrodynamics itself must play
a constitutive role. Second, the analysis can be viewed in a straightforwardway as resolving a contradiction by creating a more comprehensive point
of view. It is not that classical mechanics is falsified by phenomena such
as the Michelson-Morely experiment. But that experiment is supposed to
apply criteria for being at rest in the ether; the contradiction is that these
criteria are satisfied by systems that are in motion relative to one another.
Einstein shows, however, that our understanding of these criteria is at
fault, resting as it does on an assumption about simultaneity that no longer
has independent grounds. Properly understood, experiments like the
Michelson-Morely experiment can test for uniform velocity relative to the
ether, butin the absence of an independent criterion for simultaneity
not for motion or rest. For the experiment uses simultaneity as a criterion
for detecting differences in the speed of light arising from motion through
the ether. Under the circumstances, however, this is a piece of circularreasoning; in truth, the speed of light is providing the criterion for simul-
taneity.
Einsteins argument thus has the same dialectical structure as a famous
one of Galileos: If the traditional Aristotelian tests for the motionlessness
of the earth are passed by systems in motion relative to the earthe.g.,
by moving shipsthen we are faced by a contradiction; we arrive at a
coherent larger perspective, however, by realizing that the tests were not
tests of motionlessness after all, but tests of uniform motion. A homelier
example is the discovery of the shape of the earth from the directions of
the stars at different latitudes. It is apparently contradictory that two lines
that are both vertical should not be parallel, but it is natural and inevitable
8/10/2019 Salle, R.D. - Reconsidering Kant, Friedman, Logical Positivism & the Exact Sciences.pdf
10/21
200
if the earth is a sphere. Of course this requires us to rethink our conception
of vertical, and to recognize it as a relative term; that the earth is spher-
ical is a conception within which every local perspective on the direc-
tions of up and downfrom each of which the others are necessarily
wrongfits coherently and consistently. In just the same manner, by the
arguments of Galileo and of Einstein, the seemingly incompatible per-
spectives of different inertial frames are comprehended in a single larger
perspective. Perhaps, after such a change of perspective, the meaning of
up (or at rest or simultaneous) could never be quite the same. But
to call the new, enlarged perspective incommensurable with the old is
to miss precisely this dialectical relation between them, and therefore the
objective sense in which a deeper understanding has been achieved.
It is perhaps true in some sense that, when we translate the theory thatthe earth is a sphere into a limiting case theory on which the earth is
flat, we obtain something that would be unrecognizable to sincere believers
in a flat earth. It should nonetheless be obvious that, since the new theory
constitutes, in a perfectly straightforward sense, a dialectical resolution of
the contradictions arising from the old theory, the new theory contains
the old theorycontains, that is, every possible local point of view from
which the earth seems flatwithin a coherent whole, in which the new
conception is perfectly compatible with our grounds for believing the old.
Similarly, Einsteins theory contains Newtonian mechanics, not merely
because it makes the same predictions in the limit of small velocities
although that is also crucially importantbut because it so clearly and
coherently exhibits ones former understanding of simultaneity as an in-
complete perspective on a larger picture. It is at this conceptual level,rather than at the level of prediction or any other quantitative standard,
that the new theory has the clearest right to be called more inclusive or
comprehensive, and the resulting development cumulative.
In Einsteins revolution, then, philosophical reflection plays some of
the roles that were attributed to it by the logical positivists, and that Fried-
man has placed in a much clearer light: it provides a critical meta-scientific
perspective on the foundations of the Newtonian paradigm, from which
the inadequacy of some fundamental Newtonian concepts, and the need
for a new coordinative definition, could be clearly seen. But philosophy
also plays a more internal role, as conceptual analysis proves to be an
essential part of the construction of the new definition, and therefore of
the very creation of a new conceptual framework. And it is in this role
that philosophical analysis most directly confounds the Kuhnian view.More important, seeing it in this role suggests a broader historical per-
spective on the place of a priori principles in physics, one that reveals an
unexpected kinship between Einsteins philosophical work and that of
Newton.
8/10/2019 Salle, R.D. - Reconsidering Kant, Friedman, Logical Positivism & the Exact Sciences.pdf
11/21
, , 201
4. For details of Kants argument see Friedman 1992, chapter 3.
5. The application of the third law of motion to the attractions between the planets
was, for Huygens, one of the most questionable steps in Newtons argument for uni-versal gravitation. See Stein 1967, 179180.
Historically, Newton has not received much credit for philosophical
reflectioncertainly not from hostile critics such as Leibniz, Berkeley, or
the logical positivists, but neither from as sympathetic a reader as Kant.
To Kant, Newton seemed content to treat the laws of physics as mere
postulates without concerning himself with their a priori sources ([1786]
1911, 472). It is important to understand precisely what Kant means by
this. He did not mean, for example, that philosophy was in possession of
a priori metaphysical principles from which physical laws could be de-
rived; that was among the chief errors of dogmatic metaphysics as under-
taken by Descartes, Leibniz, and their followers. Rather, he meant, in the
spirit of the Critical philosophy, that Newtons laws did not need to be
postulated or derived: they could be shown to be conditions of the pos-
sibility of comprehending nature as a system, that is, of subsuming natureas a whole under the concepts of the understanding. Hence Kants com-
plaint against Newtons reluctance to see gravity as an immediate action
at a distance, and professed eagerness to discover the underlying causal
mechanisma profession which set Newton at variance with himself
([1786] 1911, 515), since on Kants analysis, an original attraction is a
condition of the possibility of our experience of matter.4
Moreover, it might seem implausible that Newtons revolutionary
achievements are the outcome of any sort of conceptual analysis. Univer-
sal gravitation was seen by seventeenth-century mechanical philoso-
phersnot unjustlyas a radical discontinuity with established scientific
and philosophical thinking. Of the laws of motion, Newton could claim
that they were accepted by mathematicians, meaning that, though no
one had explicitly stated them quite as Newton had, they were assumedin the work of Galileo, Huygens, Wallis, and Wren on projectile motion
and elastic collisionsindeed, this is precisely what he means when he
says that the laws are confirmed by experiments of many kinds, i.e., that
they are presupposed in the solutions of all the mechanical problems that
have actually been solved. To this extent the laws can be said to result
from a conceptual analysis of what is implicit in the work of his prede-
cessors. But to the extent that the laws are said to hold for attractions or
action at a distance as well, it seems that we have to acknowledge a genuine
novelty in Newtons conception, something not countenanced by the me-
chanical philosophy.5 And it is just this sort of novelty that a mere con-
ceptual analysis would not be expected to produce.
We know the history of Newtons attitude toward his theory of uni-
8/10/2019 Salle, R.D. - Reconsidering Kant, Friedman, Logical Positivism & the Exact Sciences.pdf
12/21
8/10/2019 Salle, R.D. - Reconsidering Kant, Friedman, Logical Positivism & the Exact Sciences.pdf
13/21
, , 203
6. As Friedman correctly points out, however, Kants remark is based on a more
specific and telling criticism: that Newtons repudiation of immediate action at a dis-tance undermines his entire project for determining the true motions in the solarsystem.
For only as immediate interactions do the mutual influences of the planets fall underthe law of action and reaction, and only under this law do they enable Newton to
determine the systems center of mass. Cf. Friedman 1992, pp. 172174.
7. Newton asserted on several occasions, with obvious reference to the mechanical
to which (along with extension) its other properties had to be reducible;
it is the property that defines how its other properties may be investigated
and understood. It is not an exhaustive account of the physical nature of
bodies, but the defining principle of a program for discovering the physical
nature of bodies, and the forces that animate them. This program is en-
capsulated in Newtons distinction between the passive principles de-
fined by the laws of motion and the active principles that cause deviation
from an inertial statebetween the laws that define inertia and force, and
the forces that those definitions enable us to discover. Rather than admit
immediate action at a distance, Newton preferred to think of such forces
as spirits or powers diffused through space. But to whatever extent
he rejected action at a distance, he was not to that extent set at variance
with himself, as Kant claimed; he would be at variance with himself onlyif he insisted that such active principles are reducible to the passive
principle of inertia, in the manner demanded by the mechanical philos-
ophy.6
In spite of its radical novelty, however, Newtons program can still be
seen to arise from a conceptual analysis of mechanics as practiced by his
mechanistic contemporaries. What enabled him to countenance powers or
forces of nature that seemed unintelligible, from the mechanistic perspec-
tive, was precisely his analysis of the intelligibility of mechanism itself
that is, his critique of prevailing notions of what made mechanical inter-
actions so preeminently intelligible in the first place. From the mechanical
view, as defended by Boyle, Huygens, and Leibniz, among others, the laws
of mechanical impact had an intelligible foundation in the nature of
bodya foundation that was lacking by definition in attractive forcesas conceived by Newton. Newton realized, however, that we have no
clearer idea of what really happens in impactof what it is in the nature
of body that explains the phenomena of collisionsthan we do of the
influence of the moon on the tides. All that we understand of impact is
precisely what is contained in the laws of impact; their intelligibility, in
other words, consists precisely in their conformity to the laws of motion.
It followed that the mechanical philosophy rested on a philosophical mis-
understanding: an intelligible interaction is not one that can be reduced
to impact, but one that satisfies the conditions under which impact itself
is intelligible.7 If Newtons theory of gravitational force is in obvious ways
8/10/2019 Salle, R.D. - Reconsidering Kant, Friedman, Logical Positivism & the Exact Sciences.pdf
14/21
204
philosophy, that we have no deeper understanding of impenetrability or extension than
we have of gravityi.e., that what he had discovered about gravity in the course of thePrincipiais at least as securely known, and at least as well understood, as the propertiesof bodies that were countenanced by the mechanists. As he remarked in the discussion
of the third Rule of Philosophizing, we know that bodies are impenetrable not byreason but by our senses ([1726] 1995, 795). Moreover, the argument from phenom-
ena will be even stronger for universal gravity than for the impenetrability of bodies,for which, of course, we have not a single experiment, and not even an observation, in
the case of the heavenly bodies ([1726] 1995, 796).
8. This discussion is not meant to gloss over the ambiguities, even the tensions, inNewtons conceptions of force and inertia, as treated in detail by McMullin (1978,
especially pp. 3347). It is meant to suggest, however, that Newtons own decision toplay down such ontological difficulties, in the presentation of the Principia, reflects
more than a reluctance to invite controversy on philosophical issues that are secondary
to the formal argument of the book (cf. McMullin 1978, 52). It also reflects the directengagement of a philosophical task that is absolutely central to the aims of the book:
to make explicit the conceptual framework implicitly shared by Newton and the me-chanical philosophers, in order to show that Newtons most controversial ideases-
pecially concerning space, time, motion, and forcemake sense within that framework.(See below and DiSalle, 2002a.)
a radical challenge to the prevailing conception of physical interaction, it
nonetheless rests on a philosophical analysis that recognizes that concep-
tion for what it truly is, i.e., an analysis that places its objective empirical
content, and its function in physical reasoning, in a clearer light. Given
this deeper understanding of force and interaction, the question what are
the forces of nature?including whether they can act otherwise than by
contactis a straightforward empirical question that can be answered by
definite procedures.8
This dialectical confrontation with prevailing conceptions, revealing
their objective content, their shortcomings, and their implications for
physics, is, I suggest, typical of Newtons philosophical interaction with
his contemporaries. Another obvious example is his solution of the frame
of the system of the world, which begins with Hypothesis I: That thecenter of the system of the world is at rest. The hypothesis is explicitly
meant as a dialectical opening, because it is the common assumption of
both sides in the controversy between the geocentric and heliocentric
views; in the analysis of the world as a dynamical system, this hypothesis
leads to the conclusion that both sides are wrong, as only the center of
gravity can be at rest ([1726] 1999, 816). Evidently this argument rests on
Newtons recognition that dynamics, as understood by himself and his
contemporaries, explicates the notion of a center of a system of bodies
in a novel way, and provides a criterion for identifying such a center that
had been entirely lacking before. A less obvious example is Newtons ar-
gument for his theory of absolute motion. As I have argued elsewhere
(DiSalle 2002a, 2002b), Newtons water bucket argument, and other
8/10/2019 Salle, R.D. - Reconsidering Kant, Friedman, Logical Positivism & the Exact Sciences.pdf
15/21
, , 205
similar arguments, are not meant to demonstrate the existence of absolute
motion (or space and time), against the relativistic theory of the Carte-
sians. Rather, they show that his conception of rotation is implicit in his
opponents own thinking; it is the very conception that they themselves
employ in their treatment of dynamical problems, above all in their vortex
theory of planetary motion. Newtons argument therefore has the char-
acter, not of an inductive argument for his hypothesis, but of a conceptual
analysis, revealing the contradictions between the ideology of the relativ-
ists, and the assumptions that implicitly guide their scientific reasoning.
Most strikingly, the analysis reveals that the most basic question posed
by Cartesian sciencewhat is the causal explanation for the celestial mo-
tions?makes sense only in a framework in which Newtons distinctions
between rotation and nonrotation, uniform motion and acceleration, aretaken for granted.
Contrary to what the Kuhnian view would lead us to expect, Newtons
philosophical criticisms are focused on a set of widely shared beliefs and
practices, rather than on a conflict between incommensurable points of
view; Newtons era was, after all, one of ascendancy rather than crisis for
the basic framework of classical mechanics. Yet Newtons philosophical
analysis of the framework was essential to its achieving maturity as a
science. This circumstance sheds some further light on the philosophical
interest that Newtonian physics had for Kant, as well as on the larger
concerns of the present paper. For the Newtonian revolution, to Kant,
manifested a philosophical revolution:
Reason . . . approaches nature in order to be taught by it: but not in
the character of a pupil who submits to everything that the teachertells him, but as an appointed judge who compels the witness to an-
swer the question which he poses. And thus even physics owes the
beneficial revolution in its way of thinking entirely to the happy
thought that, in accordance with what reason itself has originally
placed in nature, we ought to seek in nature (and not invent for it)
whatever reason has to learn from nature and could not know by
itself. In this way the study of nature first came upon the secure path
of a science, after having for many centuries done nothing but grope
in the dark. (Kant [1787] 1956, Bxiiixiv).
It is evidently no accident that such remarks reflect a proto-Kuhnian dis-
tinction between pre-paradigm science and mature science, in which
the paradigm defines a research program and a set of determinate puz-zlesin short, a framework for normal science. For Kuhn himself, as
Friedman reminds us, characterized his own general view as Kantianism
with movable categories. But for Kant this distinction is connected with
a larger philosophical problem: to explain the conceptual prerequisitesfor
8/10/2019 Salle, R.D. - Reconsidering Kant, Friedman, Logical Positivism & the Exact Sciences.pdf
16/21
206
normal science, i.e., the conceptual foundation that makes a theory ca-
pableof supporting a tradition of normal sciencei.e., capable of gener-
ating puzzles to solve, capable of generating questions that may be an-
swered by some definite procedure. And this problem was connected with
a question about the comparative failure of philosophy: how is it that
Newtonian science has achieved the kind of universal assent that has al-
ways eluded philosophy, and most of the sciences as well? Like Kuhn,
Kant did see traditional philosophy as mired in hopeless clashes of per-
spectiveinconclusive debates over fundamentals of the sort that sci-
entists engage in when they are behaving like philosophersbut hoped
that philosophy could escape this predicament by understanding how
Newtonian science had done so.
From the Kuhnian perspective, the progression from pre-paradigm sci-ence to mature science essentially involves a social decision by the relevant
group, to accept a particular form of life without further serious ques-
tion, at least until sufficiently disturbing anomalies begin to accumulate.
It is understandable that this view would recommend itself over that of,
say, Popper, who seemed to suggest that the typical practice of scientists
was to put their theories to severe tests. To the extent that philosophy of
science aims to characterize the typical practice of scientists, and in par-
ticular the relation between their empirical research and the prevailing
theory, philosophers of science could not be blamed for sharing Kuhns
assessment of Popper, that Sir Karl has characterized the entire scientific
enterprise in terms that apply only to its occasional revolutionary parts
(Kuhn 1970, 6). And Kuhn even suggested, echoing Kant, that the ability
to generate puzzles provided the demarcation criterion between scienceand pseudo-science that Popper had mistakenly sought elsewhere. But to
consider how a theory might arrive at this statewhat must characterize
the fundamental concepts of a theory, so that it might provide a founda-
tion for normal sciencelay quite beyond the scope of Kuhns concerns.
For Kant, by contrast, this was perhaps the central problem of philos-
ophy. If philosophy was to achieve the certainty of the sciences, it would
not suffice to imitate their deductive structure; as the history of philosophy
clearly showed, metaphysical systems could be just as rational as the
sciences without thereby establishing any certain results, or settling in any
way the traditional metaphysical disputes. Philosophy could be entirely
rational, in other words, without ceasing to be entirely subjective. The
difficulty lay in the essential arbitrariness of the fundamental concepts,
which tended to be incompletely understood, like those of God and thesoul, or merely invented, like that of the monad. In the exact sciences, by
contrast, deductive arguments could yield genuine knowledge because the
fundamental concepts were arrived at by construction in accord with the
forms of spatial and temporal intuitionin other words, by objective pro-
8/10/2019 Salle, R.D. - Reconsidering Kant, Friedman, Logical Positivism & the Exact Sciences.pdf
17/21
8/10/2019 Salle, R.D. - Reconsidering Kant, Friedman, Logical Positivism & the Exact Sciences.pdf
18/21
208
the views both of Kant and of the positivists, in contrast, was their sense
that the best science of their time was of philosophical interest not merely
as such, but because it had achieved something definitive as philosophy;
the point of the philosophical analysis of science, as distinct from phi-
losophy of science in our sense, was to explain how and why this was the
casenot to justify the methods or the results of science, but to show that
in the progress of science, a certain philosophical ground had been defin-
itively and irrevocably won.
Friedman, meanwhile, is content to illuminate the crucial role that phi-
losophy has played in ensuring the rationality of scientific revolutions, by
providing an intellectual context in which the prevailing paradigm can be
rationally criticized, and unorthodox ideas developed and rationally dis-
cussed. This is not simply because he is a turn-of-the-century philosopherof science addressing the late-twentieth-century doubts about the ration-
ality of science; no one has shown more sympathy and concern than Fried-
man for the broader philosophical aims of Kant and the positivists.
Rather, it is because just those philosophical aims, and just their sense of
the unique philosophical significance of their own contemporary science,
seem to embody the weaknesses of their respective philosophical outlooks.
That Newtons laws are, more than exemplary science, synthetic a priori
principles, necessary conditions of our understanding of nature, appears
to be both the most distinctively Kantian and the least credible claim in
Kants treatment of physics. The positivists treatment of relativity, on the
other hand, and of its special philosophical achievements, appealed to
epistemological notions that now seem quite naive and out of touch with
the nature and the practice of physics. Their belief that general relativityhad eliminated all metaphysical entities in favor of immediatelyobservable
relations, and that this marked a decisive break with all previous theories
of space and time, involved confusions about general relativity, the pre-
vious theories of space and time, and epistemology in general. It is with
respect to the positivists understanding of relativity, in particular, that
Kants understanding of Newtonian physics seems exemplary.
These shortcomings of Kant and the positivists, we can now see, reveal
the shortcomings in their respective views of conceptual analysis. Accord-
ing to the positivists, Kant had correctly identified the non-empirical and
formal character of the fundamental concepts of natural sciencehad rec-
ognized, that is, that they are a priori and cannot originate in experience,
because of the constitutive role that they play in our grasp of experience.
But he mistakenly viewed those concepts as founded on the a priori cate-gories of the understanding, and the laws of nature as synthetic a priori
truths about possible experience. He thus missed the point that they are
essentially meaning-constitutive or analytic, and therefore can only be
fixed by convention. But this assessment does little justice to the subtlety
8/10/2019 Salle, R.D. - Reconsidering Kant, Friedman, Logical Positivism & the Exact Sciences.pdf
19/21
, , 209
of Kants view. In regarding the constitutive principles as synthetic a
priori, Kant was not implying that they are immediately given or uncrit-
ically assumed; he had too clear a sense of their historical development,
and the efforts that had brought about their definitive expression, to take
such a view. (Kants view was clearly not Kuhnianism with a fixed par-
adigm.) On the contrary, he was recognizing that the principles had been,
and could only have been, discovered by some kind of conceptual analysis
or transcendental deduction that exposes their constitutive roleas ex-
emplified by Newtons defense of the laws of motion as the assumptions
implicit in the solution of mechanical problems. And the empirical prin-
ciples that Kant had taken to be valid a priori were overturned precisely
when a novel conceptual analysis, in novel historical and epistemic cir-
cumstances, brought some new constitutive principles to light: in the caseof Euclidean geometry, the analyses of Helmholtz and Poincare that re-
vealed the constitutive role of the principle of free mobility; in the case of
Newtonian mechanics, Einsteins analysis of the constitutive role of light-
propagation in the fixing of frames of reference. The problem with Kants
view, then, was not his failure to regard such principles as conventional
stipulations. The problem was, instead, his failure to appreciate the em-
pirical and historical contingency of the supposed conditions of the pos-
sibility of experience; he understood how such principles can be objec-
tively regarded as necessary, but missed the point that their necessity is
relativenot indeed relative to a set of stipulations, like the necessity im-
posed by the conventional rules of a game, but relative to a given body of
empirical principles and practices, of whose possibility they are the con-
ceptual conditions.The positivists, meanwhile, appreciated the contingency and the mu-
tability of constitutive principles, and gathered from Einsteins example
that fundamental concepts must be, as our empirical knowledge develops,
re-evaluated and even replaced. But they had an overly restrictive view of
the function of conceptual analysis in this process; they saw it as the de-
structive application of empiricist and antimetaphysical epistemological
criteriacriteria that have come to seem, both for scientific methodology
and for general philosophy, quite naiveand so had no account of the
construction of novel conceptual frameworks except by arbitrary stipu-
lation. The Kantian notion of the discovery of a constitutive principle
simply made no sense from their point of view.
This essay has been an effort to make sense of that notion, to free it of
the difficulties with which Kant had burdened it, and to show that it il-luminates some of the most sweeping conceptual transformations in the
history of sciencetransformations of which Kuhns notion of paradigm
shift provides only the most obscure picture. It is not surprising that Kuhn
should have doubted that scientific method, in any familiar sense, could
8/10/2019 Salle, R.D. - Reconsidering Kant, Friedman, Logical Positivism & the Exact Sciences.pdf
20/21
210
justify such a change of constitutive principles; since the basic laws of
physics cannot be understood as empirical laws, the argument for them
cannot be empirical either, but must be a kind of transcendental argument.
But such arguments give an intelligibility to conceptual transformations
that Kuhns view denies them. Moreover, they illustrate a central aspect
of Kants philosophy of science that the downfall of Euclidean geometry
and Newtonian physics has not overturned: as we have seen, the dialectical
arguments used by people such as Helmholtz, Poincare, and Einstein have
exactly this transcendental character. What is captured neither by Kants
view, nor by the conventionalism of the logical positivists, is a proper sense
of the empirical context of such transcendental arguments, a sense of the
ways in which physical conditions, and the empirical principles that we
develop in our efforts to comprehend them, can determine our conceptualresources, and limit or enlarge our conceptual horizons.
REFERENCES
DiSalle, Robert (2002a), Newtons Philosophical Analysis of Space and Time, in TheCambridge Companion to Newton.Cambridge: Cambridge University Press.
(2002b), Conventionalism and Modern Physics: A Re-Assessment, Nous 36:169200.
Earman, John (1989), World Enough and Spacetime.Cambridge, MA: M.I.T. Press.Einstein, Albert (1905), On the Electrodynamics of Moving Bodies, in Einstein et al. 1952,
3565.(1911), On the Influence of Gravitation on the Propagation of Light, in Einstein
et al. 1952, 97108.(1916), The Foundation of the General Theory of Relativity, in Einstein et al.
1952, 109164.
(1917),Uber d ie spezielle und die allgemeine Relativitatstheorie (Gemeinverstandlich),2nd ed. Braunschweig: Vieweg und Sohn.
(1922), The Meaning of Relativity.Princeton: Princeton University Press.Einstein, Albert, H. A. Lorentz, Hermann Minkowski, and Hermann Weyl (1952),The Prin-
ciple of Relativity.Translated by W. Perrett and G. B. Jeffery. New York: Dover Books.Friedman, Michael (1992), Kant and the Exact Sciences.Cambridge, MA: Harvard Univer-
sity Press.(1999), Reconsidering Logical Positivism. Cambridge: Cambridge University Press.(2002),Dynamics of Reason: The 1999 Kant Lectures at Stanford University. Chicago:
University of Chicago Press.Galilei, Galileo ([1632] 1967),Dialogue Concerning the Two Chief World SystemsPtolemaic
and Copernican. Translated by Stillman Drake. Berkeley and Los Angeles: Universityof California Press.
Hall, A. Rupert and Marie Boas Hall (eds.) (1962), Unpublished Scientific Papers of IsaacNewton.Cambridge: Cambridge University Press.
Kant, Immanuel ([1786] 1911),Metaphy sische Anfan gsgrunde der Naturwissenschaft. Volume4 ofGesammelte Schriften,Akademie Ausgabe. Berlin: Reimer.
([1787] 1956),Kritik der reinen Vernunft,2nd ed. Hamburg: Felix Meiner Verlag.Kuhn, Thomas (1962), The Structure of Scientific Revolutions. Chicago: University of Chi-
cago Press.(1970), Logic of Discovery or Psychology of Research?, in Imre Lakatos and Alan
Musgrave (eds.),Criticism and the Growth of Knowledge.Cambridge: Cambridge Uni-versity Press.
8/10/2019 Salle, R.D. - Reconsidering Kant, Friedman, Logical Positivism & the Exact Sciences.pdf
21/21
, , 211
McMullin, Ernan (1978), Newton on Matter and Activity. Notre Dame: University of NotreDame Press.
Minkowski, Hermann (1908), Space and Time, in Einstein et al. 1952, 7591.Newton, Isaac ([1692] 1995), Fourth Letter to Richard Bentley, in I. Bernard Cohen and
Richard S. Westfall (eds.), Newton: Texts, Backgrounds, Commentaries. New York:W. W. Norton, 336339.
([1726] 1999). The Principia: Mathematical Principles of Natural Philosophy, tr. I.Bernard Cohen and Anne Whitman. Berkeley and Los Angeles: University of Califor-nia Press.
([1729] 1962), The System of the World, in Florian Cajori (ed.), Sir Isaac NewtonsMathematical Principles of Natural Philosophy and His System of the World, 2 vols.Translated by Andrew Motte. Berkeley: University of California Press.
Poincare, Henri (1899), Des fondements de la geometrie: A propos dun livre de M. Rus-sell.Revue de Metaphysique et de Morale 7:251279.
(1913),The Foundations of Science: Science and Hypothesis; The Value of Science;Science and Method.Translated by George Bruce Halsted. Lancaster, PA: The SciencePress.
Reichenbach, Hans ([1927] 1957), ThePhilosophy of Space and Time. Translated by MariaReichenbach. New York: Dover Publications. Originally published as Philosophie derRaum-Zeit-Lehre(Berlin).
([1920] 1965),The Theory of Relativity and a Priori Knowledge. Translated by MariaReichenbach. Berkeley and Los Angeles: University of California Press. Originallypub-lished asRelativitatstheorie und Erkenntnis Apriori(Berlin).
Russell, Bertrand (1897),An Essay on the Foundations of Geometry.Cambridge: CambridgeUniversity Press.
Stein, Howard (1967), Newtonian Space-Time,Texas Quarterly10: 174200.Torretti, Roberto (1983), Relativity and Geometry. Oxford: Pergamon Press.
(1989),Creative Understanding.Chicago: University of Chicago Press.